A Comparison of Statistical Validity of In-Situ Hydraulic Conductivity Prediction Models of Rock Mass Inferred from Borehole Logs and Lugeon Test Data

Abstract

In-situ hydraulic conductivity is a vital property in rock engineering for jointed rock mass. Understanding its correlation with rock mass parameters is crucial for water circulation. Therefore, a study was carried out to develop statistically significant empirical relationships between hydraulic conductivity and various rock mass parameters to estimate in-situ hydraulic conductivity from Lugeon test and various rock mass parameters obtained from borehole logs.
The study initially aimed to establish a correlation between hydraulic conductivity and Rock Quality Designation (RQD). However, the outcomes were unsatisfactory, prompting further research. Later, two more robust models were developed, namely the HC-model and modified HC-model. The HC-model incorporated four rock mass parameters, including Rock Quality Designation Index, Depth Index, Gouge Content Designation Index, and Lithology Permeability Index, achieving a maximum coefficient of determination (R²) of 0.46. The modified HC-model included six parameters, encompassing fracture frequency and theoretical aperture, resulting in an improved R² of 0.69. Both models significantly outperformed RQD-alone predictions (R² < 0.10), highlighting the need for incorporating multiple rock mass parameters in predicting hydraulic conductivity due to a complex interplay of various factors. However, the effects of joint persistence and roughness are limiting in the present analysis.